Early polyadenylation signals of human papillomavirus type 31 negatively regulate capsid gene expression

Abstract

The L1 and L2 capsid genes of human papillomavirus type 31 (HPV-31) are expressed upon keratinocyte differentiation from a promoter located in the E7 open reading frame (ORF) of the early region. Late transcripts must therefore pass through and ignore the early polyadenylation sequences to use the downstream late AAUAAA element located at the end of the L1 ORF. To identify sequences which modulate downstream capsid gene expression, a variety of substitution mutations were introduced into the early polyadenylation signal and studied first in the context of polycistronic luciferase reporter constructs. Removal of the G/U-rich cleavage stimulation factor (CstF) binding sites and the degenerate cleavage and polyadenylation specificity factor binding sites, UAUAUA, had minimal effect on downstream expression as defined by luciferase activities. This is in contrast to the deletion of the HPV-31 early AAUAAA element, which resulted in a dramatic increase in downstream expression. Additional sequences within the first 800 bp of the L2 ORF were also found to negatively regulate capsid expression in luciferase assays. To determine how these mutations influence gene expression in the context of the complete HPV-31 genome, recombinant genomes were constructed that contained a substitution in the AAUAAA sequence, an inserted strong CstF binding site, an inserted simian virus 40 (SV40) late poly(A) signal, or a substitution of the 5'-most 800 nucleotides of the L2 ORF. Reductions in both transient and stable replication were observed with the recombinant genomes containing the strong CstF site or the late SV40 signal, suggesting that alterations in the strength of the upstream poly(A) signal influence expression of viral replication factors. Similarly, disruption of the L2 ORF resulted in a significant reduction in genome replication and an inability to be maintained stably. In contrast, genomes containing a substitution of the AAUAAA sequence had increased levels of transient and stable replication. Quantitation of late transcripts following keratinocyte differentiation in methylcellulose also showed a reduction in downstream capsid gene expression in lines containing genomes with the strong CstF site or the late SV40 signal mutations, while a significant increase in expression was detected in the lines with genomes lacking the AAUAAA sequence. These studies demonstrate that capsid gene expression in HPV-31 requires an inefficient early poly(A) signal which is defined primarily by the AAUAAA element as well as a major negative regulatory element located within the L2 ORF.

FIG. 1

Map of the HPV-31 genome showing the major early and late transcripts. Late messages terminate either near pA_Early or at a downstream consensus hexanucleotide element, pA_Late.

FIG. 2

Identification of HPV-31 early polyadenylation cis elements which regulate downstream gene expression. (A) Schematic of pPolyA Luc-Control reporter construct containing a cytomegalovirus (CMV) promoter driving expression of both the Renilla (Rluc) and firefly (Fluc) luciferase genes. The genes are separated by an IRES element from encephalomyocarditis virus and contain a downstream polyadenylation signal from bovine growth hormone (bGH pA). (B) The reporter pPolyA Luc-1500 contains sequences from upstream of the E5 ORF to 1,500 nt of the late coding region. The HPV-31 sequence includes the early polyadenylation signal AAUAAA (solid vertical box) and degenerative signal UAUAUA (shaded vertical box). pPolyA Luc-P1.15 and -P2.15 contain substitutions in the AAUAAA and UAUAUA elements, respectively. pPolyA Luc-C1.15, -C2.15, and -C3.15 contain substitutions which reduce the G/U content within previously defined CstF binding sites. Plasmids were transfected into LKP-31 cells as described in Materials and Methods. Luciferase activities were determined and are illustrated graphically as the ratio of relative light units (RLU) from the downstream firefly luciferase to the upstream Renilla luciferase as a percentage of that with pPolyA Luc-Control. The ratios are presented as the standard deviations from three experiments.

FIG. 3

Identification of sequences within the L2 ORF which influence downstream pPolyA Luc reporter expression. (A) Substitution mutations were introduced into the late coding region of pPolyA Luc-1500, replacing the late coding region with neutral spacer sequence obtained from the HPV-31 E1 ORF (E1). In pPolyA Luc-E15, 1,500 nt of the late coding region was replaced with E1. pPolyA Luc-EL8 and -LE8 contain substitutions in the 5′ and 3′ end of the late region, respectively. The reporters pPolyA Luc-EL4 and -LE4 contain substitutions within the 5′-most 800 nt of L2 with spacer sequence. (B) pPolyA Luc-BL8 contains a substitution within the 5′ end of L2 using spacer sequence obtained from the prokaryotic β-galactosidase gene. In pPolyA Luc-CPB, both AAUAAA and the 5′-most 800 nt of L2 have been replaced by spacer sequence. Reporter plasmids were transfected into LKP-31 cells as described in Materials and Methods. Activities are illustrated graphically as a ratio of RLUs from firefly luciferase to Renilla luciferase and presented as a percentage of the pPolyA Luc-Control value. The ratios are presented as the standard deviations from three experiments with the exception of the pPolyA Luc-EL4 data, which come from a single experiment.

FIG. 4

Transient-replication assays using HPV-31 genomes containing mutations within the early polyadenylation signal. (A) Transient-replication assays were completed using viral DNA that was released from the bacterial vector, unimolecularly ligated, and transfected into SCC13 cells. After 5 days, low-molecular-mass DNA was isolated through Hirt extraction. DNA was digested with DpnI to remove methylated input DNA, and replication levels were determined through Southern analysis using a ³²P-labeled HPV-31 genomic DNA fragment (lanes 5 to 9). Lanes 1 to 4 are DNA standards (Std) consisting of linearized wild-type (WT) HPV-31 DNA representing 500, 25, 2.5, and 0.5 pg, respectively. Replication was restored to wild-type levels upon cotransfection with E1 and E2 expression vectors (lanes 10 to 14). (B) Levels of replication from lanes 5 to 9 have been quantitated through phosphorimage analysis and are presented relative to the wild-type value. The data are presented as the standard deviation from multiple experiments. (C) Mutations were introduced into p599-31WT which contains the HPV-31 genome in pBR322 as described within the Material and Methods. Plasmid p599-31CP contains a substitution in the early AAUAAA element, while p599-31LD contains the late strong CstF binding site 40 nt downstream of AAUAAA. Plasmid p599-31SV contains the complete SV40 late polyadenylation signal in place of the early signal. Plasmid p599-31BL contains a substitution in the 5′-most 800 nt of the L2 ORF using a neutral spacer sequence obtained from the β-galactosidase gene.

FIG. 5

Southern blot analysis of NHK lines containing recombinant wild-type and mutant HPV-31 genomes. NHK lines containing recombinant HPV-31 genomes were established as described in Materials and Methods. Southern blot analysis was completed using 10.0 μg of total DNA isolated from low-passage 31WT.1, 31CP.1, 31LD.1, and 31SV.1 cells lines. DNA was digested with either DpnI, to remove residual input DNA (lanes 9 to 12), or DpnI and BlpI, to linearize HPV-31 DNA, allowing quantitation (lanes 5 to 8). Specific DNAs were detected with ³²P-labeled HPV-31 genomic DNA fragment. Several species of DNA are observed, including open circle (oc), linear (ln), and supercoiled (sc). Lanes 1 to 4 are genome copy number controls representing 100, 50, 25, and 5 copies per cell, respectively.

FIG. 6

Quantitation through real-time RT-PCR of E1∧E4-containing transcript levels within undifferentiated NHK cells containing recombinant HPV-31 mutants. (A) The concentration of E1∧E4 transcripts was determined through real-time RT-PCR using 2.0 μg of total RNA isolated from NHK lines containing recombinant genomes. Total RNA was isolated from lines placed in semisolid medium for 2 and 24 h. Quantitation was completed through SYBR green I dye incorporation into the amplicon, followed by quantitation of fluorescence. The RNA standard was synthesized through in vitro transcription reactions from an E1∧E4,L1 cDNA and amplified using the same primer set. Levels of E1∧E4-containing transcripts were determined from a standard curve (Stds). Similar results were seen in multiple experiments. (B) Quantitation was completed using a primer set which spans E1∧E4 using a donor splice site at nt 877 and an acceptor at nt 3295.

FIG. 7

Quantitation through real-time RT-PCR of E4∧L1-containing transcript levels from differentiated NHK lines containing recombinant HPV-31 genomes. (A) Real-time RT-PCR was completed using 2.0 μg of total RNA isolated from cells placed in semisolid medium for 24 h. The RNA standard was synthesized through an in vitro transcription reaction using the E1∧E4,L1 cDNA. The concentration of E4∧L1 transcripts was determined through a standard curve (Stds). Similar results were seen in multiple experiments. (B) Quantitation was completed with a primer set that spans E4∧L1 using a splice donor at nt 3590 and a splice acceptor at nt 5552.

FIG. 8

Relative readthrough activity of HPV-31 wild-type and mutant polyadenylation signals in the context of the complete genome within differentiated NHKs. The ratio of E4∧L1- to E1∧E4-containing transcripts, or readthrough activity, is presented relative to the wild-type lines. The generation of stable lines was repeated twice using two different foreskin donors, NHK.1 and NHK.2, to ensure reproducibility.